Skip to main content
Article
Online Supervised Learning of Non-Understanding Recovery Policies
SLT-2006, Palm Beach, Aruba.
  • Dan Bohus, Carnegie Mellon University
  • Brian Langner, Carnegie Mellon University
  • Antoine Raux, Carnegie Mellon University
  • Alan Black, Carnegie Mellon University
  • Maxine Eskenazi, Carnegie Mellon University
  • Alexander I Rudnicky, Carnegie Mellon University
Date of Original Version
1-1-2006
Type
Conference Proceeding
Abstract or Description

Spoken dialog systems typically use a limited number of non- understanding recovery strategies and simple heuristic policies to engage them (e.g. first ask user to repeat, then give help, then transfer to an operator). We propose a supervised, online method for learning a non-understanding recovery policy over a large set of recovery strategies. The approach consists of two steps: first, we construct runtime estimates for the likelihood of success of each recovery strategy, and then we use these estimates to construct a policy. An experiment with a publicly available spoken dialog system shows that the learned policy produced a 12.5% relative improvement in the non-understanding recovery rate.

Citation Information
Dan Bohus, Brian Langner, Antoine Raux, Alan Black, et al.. "Online Supervised Learning of Non-Understanding Recovery Policies" SLT-2006, Palm Beach, Aruba. (2006)
Available at: http://works.bepress.com/alexander_rudnicky/72/